Electric bicycle battery overheating safety monitoring device

Abstract

The utility model discloses an overheat safety monitoring device for a battery of an electric bicycle. The overheat safety monitoring device comprises an input module, a control processing module and an output module, the input module comprises an infrared thermal imager and a visible light acquisition device, and the infrared thermal imager and the visible light acquisition device are both connected with the control processing module and transmit data to the control processing module; the control processing module is provided with a Raspberry Pi development board and a server, and the Raspberry Pi development board is respectively connected with the infrared thermal imager and the visible light acquisition device and is used for receiving data acquired by the infrared thermal imager and the visible light acquisition device and sending the data to the connected server; the output module comprises an alarm and a management terminal, and the alarm and the management terminal are both connected with the server, receive data sent by the server, give an alarm according to the data and send the data to the management terminal.

Description

TECHNICAL FIELD

[0001] The utility model relates to battery protection technical field especially relates to a kind of electric bicycle battery overheating safety monitoring device. BACKGROUND

[0002] The popularity of existing electric bicycles greatly facilitates people's daily travel. Compared with traditional bicycles, electric bicycles do not require physical effort, providing considerable convenience and timeliness for many people who need to walk. Electric bicycles play an important role in relieving urban traffic pressure. With the increase in the number of urban motor vehicles, traffic congestion is becoming more and more serious. Electric bicycles, with their small size and strong flexibility, can easily navigate through the complex urban roads, effectively reducing the likelihood of traffic congestion. Many cities have begun to encourage residents to use electric bicycles as their daily transportation, further improving the convenience of electric bicycles through measures such as planning dedicated lanes and setting up charging stations. This change not only improves travel efficiency, but also improves the overall condition of urban traffic.

[0003] China is the world's largest user of electric bicycles. According to statistics from the Ministry of Industry and Information Technology, the social ownership of electric bicycles in China exceeded 350 million by the end of 2023. Electric bicycles have gradually become an important basic transportation tool for consumers' daily short trips, with a large market size and continued growth. This data shows that electric bicycles have become the main mode of transportation for short trips for Chinese residents, especially in first-tier cities, where the popularity of electric bicycles is very high.

[0004] Due to its flexibility, environmental friendliness and convenience, electric bicycles can meet many people's needs for short-distance travel, such as picking up and dropping off children from school, going out to shop, etc. The social ownership of electric bicycles is also increasing.

[0005] Electric bicycles refer to electromechanical integrated personal transportation tools that install electric motors, controllers, batteries, handlebar brakes, and display instrument systems on the basis of ordinary bicycles using batteries as auxiliary energy. Electric bicycle batteries are diverse, each with its unique characteristics and application scenarios.

[0006] (1) Lead-acid batteries: the best choice in terms of price, but they have a relatively short cycle life, are relatively heavy, and have some environmental problems.

[0007] (2) Lithium batteries: one of the most popular electric vehicle batteries on the market today, with high energy density, lightweight, strong power storage capacity, long cycle life, and excellent environmental adaptability.

[0008] (3) Nickel-metal hydride battery: It is a transitional battery between lead-acid battery and lithium-ion battery, with low memory effect, strong overcharge resistance, up to three thousand times of cycle life and relatively good energy density. But its high temperature resistance is not good, the charging efficiency is low, the self-discharge rate is high, and the material cost is high. It is less used as an electric bicycle battery.

[0009] Lithium battery is light and has strong endurance, which is the best choice for electric bicycles. Lithium battery has become the battery of most electric bicycles.

[0010] The fire incident caused by the spontaneous combustion of electric bicycle battery has attracted widespread social attention

[0011] With the popularity of electric bicycles, especially the use of lithium batteries in electric bicycles, fire accidents caused by battery spontaneous combustion occur frequently. Therefore, this study carried out an investigation:

[0012] Table 1 Investigation of fire incidents caused by spontaneous combustion of electric bicycle battery in recent years

[0013]

[0014] Main causes of electric bicycle fire

[0015] The main cause of electric bicycle fire is battery, and improper charging and quality of battery can cause spontaneous combustion, which can cause fire. The specific analysis is as follows:

[0016] (1) Insufficient capacity of lithium battery negative electrode: When the capacity of the negative electrode part of the positive electrode part of the lithium battery is insufficient, the lithium atoms generated during charging cannot be inserted into the interlayer structure of the negative electrode graphite, and will be deposited on the surface of the negative electrode, forming a crystal. Long-term formation of crystals in lithium batteries can cause short circuits, and when the battery cell discharges sharply, it can generate a lot of heat and burn the separator. High temperature can make the electrolyte decompose into gas, and when the pressure is too large, the battery cell will explode.

[0017] (2) High water content: During charging, water can react with lithium to form lithium oxide, causing the capacity of the battery cell to be lost, making the battery cell prone to overcharging and generating gas. The decomposition of water can cause the voltage to be low, and it is easy to decompose and generate gas during charging. This series of generated gas will increase the internal pressure of the battery cell, and when the shell of the battery cell cannot withstand it, the battery cell will explode.

[0018] (3) Internal short circuit: Internal short circuit causes large current discharge, generates a large amount of heat, burns the separator, causes more short circuit phenomenon, and makes the electrolyte decompose into gas, causing the internal pressure to be too large, thereby causing the battery cell to explode.

[0019] (4) Long-term overcharge of lithium battery: when the battery is overcharged, the lithium over-release of the positive electrode will change the structure of the positive electrode, and too much lithium released is also difficult to insert into the negative electrode, which is easy to cause the phenomenon that lithium ions are not embedded in the negative electrode material on the negative electrode side, but are precipitated on the negative electrode surface in the form of metallic lithium. When the voltage reaches 4.5V or more, the electrolyte will decompose to produce a large amount of gas, causing explosion.

[0020] The market electric bicycle battery is mainly lithium battery, and the reason for causing spontaneous combustion due to overheating is mainly related to thermal runaway of the battery. During the thermal runaway process, the reactions inside the battery mainly include SEI film decomposition, negative electrode and electrolyte reaction, positive electrode and electrolyte reaction, separator melting, electrolyte decomposition and adhesive decomposition. The SEI film is a thin film formed on the surface of the graphite negative electrode during the first cycle, and when the battery temperature reaches 69℃, the SEI film will decompose and release heat, and release CO2 and other gases. When the temperature rises to about 90℃, the Li in the negative electrode begins to react with the electrolyte due to the loss of protection of the SEI film. If the temperature exceeds 180℃, the positive electrode of the battery will decompose and release a large amount of heat, and release oxygen. When the temperature exceeds 180-200℃, the electrolyte and adhesive inside the battery will successively decompose.

[0021] Therefore, the principle of thermal runaway of the battery shows that when the temperature exceeds a certain threshold (65℃), the battery begins to release heat, and continuous release leads to spontaneous combustion. SUMMARY

[0022] To solve the above technical problems, the purpose of the utility model is to provide an electric bicycle battery overheating safety monitoring device.

[0023] The purpose of the utility model is realized by the following technical scheme:

[0024] An electric bicycle battery overheating safety monitoring device, comprising:

[0025] The input module, the control processing module and the output module are included.

[0026] The input module includes an infrared thermal imager and a visible light acquisition device, and the infrared thermal imager and the visible light acquisition device are connected with the control processing module and transmit data to the control processing module; the control processing module is provided with a Raspberry Pi development board and a server, the Raspberry Pi development board is connected with the infrared thermal imager and the visible light acquisition device respectively, used for receiving the data collected by the infrared thermal imager and the visible light acquisition device, and sending the data to the connected server end; the output module includes an alarm and a management terminal, and the alarm and the management terminal are connected with the server and receive the data sent by the server, and alarm and send the management terminal according to the data.

[0027] Compared with the prior art, one or more embodiments of the utility model can have the following advantages:

[0028] The battery temperature is detected by using the infrared thermal imager, the alarm can be used to tell the relevant personnel or the user in an emergency, and real-time on-site data can be provided for the monitoring personnel. BRIEF DESCRIPTION OF DRAWINGS

[0029] Figure 1 It is a schematic diagram of the electric bicycle battery overheating safety monitoring device module;

[0030] Figure 2 It is an appearance structure diagram of the electric bicycle battery overheating safety monitoring device;

[0031] Figure 3 It is a model training and application diagram;

[0032] Figure 4 It is a schematic diagram of the model training loss function changing with the determination accuracy and the training period. DETAILED DESCRIPTION

[0033] In order to make the purpose, technical scheme and advantages of the utility model more clear, the following will combine examples and drawings to make further detailed description on the utility model embodiment.

[0034] As Figures 1-3 shown, it is an electric bicycle battery overheating safety monitoring device, which comprises an input module, a control processing module and an output module;

[0035] The input module comprises an infrared thermal imager and a visible light collecting device, the infrared thermal imager and the visible light collecting device are connected with the control processing module and transmit data to the control processing module; the control processing module is provided with a Raspberry Pi development board and a server, the Raspberry Pi development board is connected with the infrared thermal imager and the visible light collecting device respectively, is used for receiving the data collected by the infrared thermal imager and the visible light collecting device, and sends the data to the connected server end; the output module comprises an alarm and a management terminal, the alarm and the management terminal are connected with the server and receive the data sent by the server, alarm according to the data and send the management terminal, as shown in table 1:

[0036] Table 1

[0037]

[0038] The infrared thermal imager adopts an infrared thermal imaging sensor, and real-time battery surface temperature data are collected; the visible light collecting device adopts a visible light camera, which is used for collecting the external image of the electric bicycle; as shown in table 2, the thermal imager parameters are as follows:

[0039] Table 2

[0040]

[0041] The device also includes an electric bicycle recognition module and an abnormal temperature recognition module, which uses a Yolov11n target detection model; the model training module is installed on the alarm.

[0042] The infrared thermal imager and the visible light camera are arranged in parallel to capture the same picture and lock the electric bicycle.

[0043] The device also includes a labelimg image annotation module, which includes an electric bicycle recognition unit, a bicycle recognition unit, an electric bicycle seat recognition unit, an electric bicycle tail storage box recognition unit, and a helmet recognition unit.

[0044] The electric bicycle is mainly lithium battery, when the temperature of lithium battery exceeds a certain threshold (65℃), the battery begins to release heat, and continuous release of heat leads to spontaneous combustion. Therefore, the safety temperature threshold of this embodiment is set to 65℃. When the battery temperature exceeds the set safety threshold, the monitoring guard will immediately issue an audible and visual alarm to remind users and management personnel of the potential overheating risk and take timely measures to avoid possible accidents.

[0045] The electric bicycle battery overheating identification model in use: a certain number of electric vehicle image samples are collected using a camera, and they are labeled; next, data preprocessing is performed to reduce noise interference on training, increase the number of training samples, and improve the generalization ability of the model; then, the yolo network is used to train these training samples, and the parameters of the pre-trained model are loaded for further training. The performance of the model is evaluated by the test set data that did not participate in the training.

[0046] Since the electric bicycle battery is generally installed under the seat cushion, a mobile phone is used to collect electric bicycle seat cushion photos to ensure sample diversity, including different angles, lighting conditions and backgrounds. Electric bicycles are often on outdoor streets, parking sheds, subway stations, etc. Therefore, data collection is carried out at the above-mentioned locations.

[0047] The initial temperature of the battery of the electric bicycle that causes spontaneous combustion is 65℃, and the normal temperature is 25℃ to 64℃. Due to safety issues, the heated object for temperature data collection cannot use objects containing chemicals to avoid injury, therefore, water is used instead of the battery of the electric bicycle, and secondly, a thermal imager is used to collect data on the normal and overheated (exceeding 65℃) batteries under the seat cushion of the electric bicycle.

[0048] To achieve electric bicycle battery identification and recognition of battery overheating images collected by thermal imagers, the yolov11n model is selected. This model can detect multiple categories of objects and adapt to different environments and times. It can also detect targets in real time, with faster inference speed, making it suitable for handling rapidly changing battery heating scenarios for electric bicycles.

[0049] (1) High performance: yolov11n generally strikes a good balance between speed and accuracy, making it suitable for real-time target detection applications.

[0050] (2) Lightweight: yolov11n is optimized in terms of model size and computational requirements, making it suitable for deployment on edge devices and mobile devices.

[0051] (3) Extensibility and ease of use: yolov11n provides easy-to-use APIs, allowing developers to easily integrate and use it for target detection.

[0052] (4) Support for multiple data formats: Supports different data formats such as COCO, VOC, etc., making it convenient for users to train and test.

[0053] The yolov11n model will be trained using the open-source framework PyTorch and deployed on the terminal buzzer of the electric bicycle parking shed.

[0054] To train the yolov11n model, data is collected in two aspects:

[0055] Electric bicycle seat cushion image data: Since electric bicycle batteries are generally installed under the seat cushion, collect electric bicycle seat cushion photos to ensure sample diversity, including different angles, lighting conditions, and backgrounds. The training of the yolo model needs to adapt to different situations, and the angle of the data taken needs to be adjusted slightly in different angles and distances.

[0056] Electric bicycle seat cushion appearance data collection

[0057] Go to school or community charging piles, subway entrances, and other main electric bicycle parking places to collect electric bicycle seat cushion data, and finally obtain about 500 photos. Since the battery overheating safety monitor is deployed above the charging shed, street, etc., data is collected from above the electric bicycle.

[0058] Electric bicycle battery normal and abnormal temperature data collection

[0059] Considering the temperature of major cities in China, data is collected under two conditions: normal and abnormal temperature of electric bicycle batteries in autumn and winter (temperature 18℃) and spring and summer (temperature 28℃). Water is used to simulate the battery and is installed under the electric bicycle seat cushion.

[0060] The identification processing data of the electric bicycle needs to use the preprocessing model, and the labelimg image annotation tool is used.

[0061] Labelimg is mainly used for image annotation in the field of machine learning and computer vision, and it supports multiple annotation formats, including the format of yolo model.

[0062] The use of Labelimg is as follows:

[0063] Enter the labelimg command in the command line or terminal to start the LabelImg software, and enter the interface style of labelimg as follows:

[0064] Define 6 labels, including "eseat", "ebike", "box", "bseat", "bike",

[0065] "toukui", which is convenient for classification of yolov11n model.

[0066] eseat: that is, the seat of the electric bicycle, used to train the model to identify the seat of the electric bicycle, so that the model can automatically find the battery under the seat.

[0067] ebike: that is, the electric bicycle, used to identify the electric bicycle, so that the model can automatically lock the electric bicycle and make the model better understand the semantic information that the seat of the electric bicycle is located inside the electric bicycle.

[0068] eike: that is, the bicycle, used to help the model distinguish between electric bicycles and bicycles.

[0069] eseat: that is, the seat of the bicycle. Similarly, it helps the model to distinguish between electric bicycles and bicycles.

[0070] If there are only the above 4 labels, the model is easy to misidentify the storage box and helmet as the seat in the test process, resulting in inaccurate subsequent tasks, so two class labels are added as follows:

[0071] box: storage box at the tail of the electric bicycle.

[0072] toukui: helmet.

[0073] Yolov11n is a lightweight target detection model with low computing requirements, fast inference speed, and the ability to efficiently deploy on embedded devices and mobile terminals, making it an ideal choice for real-time detection tasks.

[0074] The accuracy of the identification degree of the electric bicycle needs to be very high, so the yolov11n model is used.

[0075] In training, a pre-trained model on coco128 target detection dataset is imported first, since the model is pre-trained on a large dataset, it has good image feature extraction capability, so it can help the model adapt to the auxiliary identification of electric bicycle parts task faster, and improve its classification performance in the training process.

[0076] The training results show that, as shown in Figure 4 The loss function decreases with the change of the determination accuracy and the training period, which reflects the gradual improvement of the detection accuracy of the model during training, and finally the performance of the model tends to converge.

[0077] To prove the superiority, the accuracy of the model in detecting electric bicycle parts in the validation set is tested; as shown in Table 3, the accuracy of the electric bicycle parts is:

[0078] Table 3

[0079]

[0080] According to the table, it can be concluded that the accuracy rate returned by the model is relatively high, and the ability to distinguish right and wrong is stable, which can provide good performance.

[0081] The device provided by the embodiment is arranged on the side of the public place, which is consistent with the monitoring camera. The infrared thermal imager is parallel to the visible light camera, and the purpose is to shoot the same picture, which is suitable for locking the electric bicycle. The detection principle of the infrared thermal imager is to compare the temperature of the battery under the seat of the electric bicycle detected by the visible light camera with the temperature of other batteries. If the temperature of the battery under the seat of the electric bicycle is higher than that of other batteries, it is considered to be dangerous. If it is lower than or equal to the temperature of other batteries, it is considered to be normal. In order to improve the accuracy of the detection temperature, when the battery temperature detected by the infrared thermal imager is higher than 65 DEG C, it is also considered to be dangerous, and when the battery temperature is lower than 65 DEG C, it is considered to be not dangerous.

[0082] Although the embodiments disclosed in the utility model are as above, the content described is only for the purpose of understanding the embodiments adopted by the utility model, and is not used to limit the utility model. Any person skilled in the art of the utility model can make any modification and change in the form and details without departing from the spirit and scope of the utility model disclosed by the utility model. The patent protection scope of the utility model shall be subject to the scope defined by the attached claims.

Claims

1. An electric bicycle battery overheating safety monitoring device, characterized in that, It includes an input module, a control processing module, and an output module; The input module includes an infrared thermal imager and a visible light acquisition device, both of which are connected to the control processing module and transmit data to it. The control processing module includes a Raspberry Pi development board and a server. The Raspberry Pi development board is connected to both the infrared thermal imager and the visible light acquisition device, receiving data from them and sending it to the connected server. The output module includes an alarm and a management terminal, both connected to the server and receiving data from it. The alarm and management terminal then trigger an alarm and send a response to the management terminal.

2. The electric bicycle battery overheating safety monitoring device according to claim 1, characterized in that, The infrared thermal imager uses an infrared thermal imaging sensor to collect real-time battery surface temperature data; the visible light acquisition device uses a visible light camera to acquire images of the electric bicycle's exterior.

3. The electric bicycle battery overheating safety monitoring device according to claim 1, characterized in that, The device also includes an electric bicycle battery identification module and an abnormal temperature identification module, which adopt the YOLOv11n target detection model; the YOLOv11n target detection model is installed on the alarm.

4. The electric bicycle battery overheating safety monitoring device according to claim 2, characterized in that, The infrared thermal imager and the visible light camera are set in parallel to capture the same image and lock the electric bicycle.

5. The electric bicycle battery overheating safety monitoring device according to claim 1, characterized in that, The device also includes a labelimg image annotation module, which includes an electric bicycle recognition unit, a bicycle recognition unit, an electric bicycle seat recognition unit, an electric bicycle rear storage box recognition unit, and a helmet recognition unit.

Images